[Users] Q: Saving multiple outputs at the same time

Helene Huck helene.huck at dgb-group.com
Wed Mar 5 17:32:00 CET 2008


Dear Antonio,

You are indeed perfectly right, there is a way for any user to 'hack'
the job *.par file to compute all the outputs of a single attribute at
once; you will then have to edit your parameter file this way:

The attached file energy_1output.par is an example of the initial
parameter file you get by creating a normal batch program for only one
output, you will use it as a base for the more elaborated one.

What you have to do is copy paste the attribute as many times as you
want and change the output=X at the end of each definition string, you
can change the name of the attribute BUT do NOT change any other parameter.

then at the end of the file you will notice these lines:
Output.1.Attributes.MaxNrKeys: 1
Output.1.Attributes.0: 2

you can change that into :
Output.1.Attributes.MaxNrKeys: 3
Output.1.Attributes.0: 2
Output.1.Attributes.1: 3
Output.1.Attributes.2: 4

to select all the attributes you created with the previous commands.
You have a good example in energy_3outputs.par

This will create an output cube with multiple components. Presently
there is no GUI support in OpendTect for multiple components cubes but
we do see the benefits of having this and we suggest to add this feature
to the wish list. Since you are familiar with our open source business
model I trust you also realize that the fastest and surest way to get
something from the wish list implemented is by sponsoring the
development. Please contact Paul de Groot if you wish to receive a proposal.

This sums up the way you can achieve your goal ( computing multiple
outputs of a single attribute in one go ) from a user side; for those
who develop plugins to opendTect there is a (much friendlier) way to do
the same with a few lines of code, I will develop the subject further in
a post to the developers list.

best regards,
Hélène




Corrao Antonio wrote:

> Hi,
>
> I am currently developing a couple of plugins that share some
> peculiarities:
>
>  a) they are computationally intensive (comparable to steering+++, or
> worse),
>
> b) they compute multiple outputs,
>
> c) such outputs are interdependent, but not in a trivial way (i.e.
> they can not be derived from other outputs alone, maybe in a
> post-processing stage),
>
> d) all of the outputs need to be generated and saved.
>
> A meaningful example would be P and S impedance inversion, volumes
> cross equalization, and more.
>
>  Now, the only way to create a seismic volume for each output, is to
> re-run the job with the same attribute definitions, apart from
> changing the output index.
>
> If I have N desired outputs, I need to launch N jobs whose cost is
> pretty much the same: suppose each job takes one week, and then you
> realize that it might be useful to generate multiple volumes at the
> same time, if possible.
>
>  Of course expanding the cluster will help minimize the problem, but
> it still would be preferable to avoid redundant computations.   
>
>  *The steering cube computation solves the same problem by generating
> a multi-dimensional output in one go: would it be sufficient to hack
> the job *.par file to achieve a similar result even for non-steering
> volumes? *
>
> I understand that without a GUI support such approach would be error
> prone, but if it works then it’s worth knowing.
>
> [BTW, a multicomponent volume would then need some custom plugin do
> extract each component, but that’s a minor issue; also export to segy
> might need a change to allow an additional ‘component’ selection].
>
> Best Regards,
>
> *Antonio Corrao*
>
> Senior Geophysicist
>
> GEOS dept. - Eni E&P Division
>
> V. Emilia 1 - 20097 San Donato Milanese - Italy
>
> Tel:+39.02.520.61150 - Fax:+39.02.520.63897 – mailto:antonio.corrao at eni.it
>
> _____________________________________________________________________________
>
> Eni S.p.A. - Sede legale in Roma, Piazzale Enrico Mattei 1, 00144 Roma
> - Cap.sociale 4.005.358.876,00 i.v.
>
> Registro Imprese di Roma - Codice Fisc. 00484960588 - P.IVA
> 00905811006 - R.E.A. Roma n.756453


-- 
-- Helene Huck
-- Geophysical software engineer
-- dGB Earth Sciences B.V.
-- Nijverheidstraat 11-2, 7511 JM Enschede, The Netherlands
-- mailto: helene.huck at dgb-group.com, http://www.dgb-group.com
-- Tel: +31 53 4315155 , Fax: +31 53 4315104

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: energy_1output.par
URL: <http://lists.opendtect.org/pipermail/users/attachments/20080305/515e1c04/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: energy_3outputs.par
URL: <http://lists.opendtect.org/pipermail/users/attachments/20080305/515e1c04/attachment-0001.ksh>


More information about the Users mailing list